Abstract: Tobacco silk production is an important process of cigarette production. The quality of tobacco silk production plays a decisive role in the quality of cigarettes, and the moisture content of tobacco is the most important quality index. The tobacco leaf silk making process includes the production processes of loose moisture recovery, leaf embellishment, silk drying, material storage and so on. The whole process controls the water content of cut tobacco by adding water, feeding, drying and other processes. In the production practice, adding water, feeding drying as well as volume, temperature and other indicators are controlled through the industrial control system combined with the experience of the operators. A large number of practices show that environmental temperature and humidity have a significant impact on the setting and control of process indicators. Through the analysis of historical production data and field investigation, it is found that in actual production, it often depends on the experience of the operators to adjust process indicators to deal with the impact of environmental temperature and humidity on the moisture content of cut tobacco. Through the establishment of multiple regression model to reflect the relationship between environmental temperature and humidity on the moisture content of cut tobacco and the set value of process indicators, the dependence on the experience of operators can be minimized to further improve the automation rate of tobacco silk process, improve the production efficiency, and realize the intelligent control of tobacco silk production process.

Key words: regression model, tobacco leaf silk, water control, large lag control